eG Help

The Web Intelligence Server Cache significantly improves load time of documents by servicing requests using documents that have already been viewed. However, if this cache is sized poorly, it will not be able to hold many documents, forcing requests to be routed to the Web Intelligence Server for processing; this is bound to increase processing overheads and consequently, the load time of documents. To avoid this, administrators should track cache usage and tune the cache size based on how it is used. This is what the BobiWISCacheTest test helps administrators do.

This test tracks the load on the cache vis-à-vis its size and notifies administrators if the cache does not have enough memory resources to cater to the demand for documents. Additionally, the test also provides effective pointers to how cache size can be increased and usage can be optimized.

Measurement	Description	Measurement Unit	Interpretation
cacheSizeKB	Indicates the amount of data stored in the cache.	KB
outOfDateDocsInCacheRate	Indicates the rate at which obsolete documents are deleted from the cache.	Docs/Sec	A higher rate is generally desired as it releases cache memory for use by new documents, and thus improves the cache's request handling ability.
maxCacheReachedRate	Indicates the rate at which the server cache size reached the maximum size limit allowed on the server.	KB/sec	A high rate indicates that the cache is getting filled up very often. This implies that the cache may require a resizing, as the load on the server is high.
docsOpenedFromCache	Indicates the number of documents for which the last request result has been directly read from the cache.	Number	By comparing these two measures at a given point in time, you can measure how effectively the cache has been servicing document requests to the Web Intelligence server. If this comparison reveals ineffective cache usage, then you may want to check the cache size to see if the poor size of the cache is responsible for this.
noOfDocs	Indicates the total number of documents that are currently open on the server.	Number
pctDocsOpenedFromCache	Indicates the percentage of documents opened from the cache.	Percent	Ideally, the value of this measure should be high. A low value is indicative of inoptimal cache usage. You may want to check the cache size in this case to see if the poor size of the cache is responsible for this.
noOfDocsWithSwapRequests	Indicates the number of documents for which a clean up thread has scheduled swap requests.	Number
pctSwappedDocuments	Indicates the percentage of documents that have been swapped by swap requests.	Percent	Swapping ensures that inactive documents in the cache are swapped to the hard disk, so that the memory they have been hogging up until then can be released for use of subsequent documents. A high percentage of documents swapped is good news, as it increases unused cache memory and enables the cache to hold more documents, which translates into more requests served.